A two-condition Kirkwood-Dirac test as the dividing line
A paper by Jonathan J. Thio and colleagues at Cambridge's Cavendish Laboratory in Physical Review Letters shows that a resource long labelled a 'magic state' in quantum computing only promises an edge over classical machines when the Kirkwood-Dirac distribution turns negative; where the distribution stays nonnegative, a classical computer can follow the same circuit efficiently, and the authors back the finding with a published classical simulation.[1]
The result is a theoretical proof paired with a classical simulation, and even before any hardware has been taken apart and rebuilt, its weight replaces the question of qubit count with the question of which flavour of magic and what its Kirkwood-Dirac signature is.[1]
The bar for a useful-qubit demonstration of advantage is now higher
Carrying the Cavendish result into the advantage accounting, a quantum-advantage demonstration must now prove two things together: that it prepared the right magic resource states and that the Kirkwood-Dirac distribution stays negative throughout the circuit; if either condition fails, an ordinary laptop running the same computation efficiently provides the comparison run.[1]
The same Kirkwood-Dirac test can also be read as a levelling that a subtler classical method might partly overtake; in the authors' framework this reading does not invalidate the result but sharpens the test itself by narrowing the restricted class further, and the difference can only be closed by independent review and work on the published classical simulation.[1]
Where to look next
A concrete signal to watch is that over the next three to six months, any new quantum-advantage claim should report the circuit's Kirkwood-Dirac negativity profile separately and compare it with a peer-reviewed classical simulation; a claim without those two additions is left comparable to a case where an ordinary laptop merely needed longer running time.[1]